package jalview.analysis;

import static org.junit.Assert.assertEquals;
import static org.junit.Assert.assertNull;

import org.junit.Before;
import org.junit.Test;

import jalview.datamodel.Alignment;
import jalview.datamodel.AlignmentAnnotation;
import jalview.datamodel.Sequence;
import jalview.datamodel.SequenceI;

public class ParsePropertiesTest
{

  private Alignment al;

  private ParseProperties pp;

  /**
   * Construct an alignment with 4 sequences with varying description format
   */
  @Before
  public void setUp()
  {
    SequenceI[] seqs = new SequenceI[]
    { new Sequence("sq1", "THISISAPLACEHOLDER"),
        new Sequence("sq2", "THISISAPLACEHOLDER"),
        new Sequence("sq3", "THISISAPLACEHOLDER"),
        new Sequence("sq4", "THISISAPLACEHOLDER") };
    seqs[0].setDescription("1 mydescription1");
    seqs[1].setDescription("mydescription2");
    seqs[2].setDescription("2. 0.1 mydescription+3");
    seqs[3].setDescription("3 0.01 mydescription4");
    al = new Alignment(seqs);

    pp = new ParseProperties(al);

  }

  /**
   * Test with a description pattern that matches any string ending in one or
   * more 'number characters' (0-9+.), i.e. greedily matches any trailing
   * numeric part of the string
   */
  @Test
  public void testGetScoresFromDescription()
  {
    // TODO - test the regex actually used by Jalview?
    // \\W*([-+eE0-9.]+)
    // see AlignFrame.extractScores_actionPerformed
    String regex = ".*([-0-9.+]+)";
    final int count = pp.getScoresFromDescription("my Score",
            "my Score Description", regex, true);
    System.out.println("Matched " + count + " for " + regex);
    assertEquals(4, count);

    /*
     * Verify values 1/2/3/4 have been parsed from sequence descriptions
     */
    AlignmentAnnotation[] anns = al.getSequenceAt(0).getAnnotation();
    assertEquals(1, anns.length);
    assertEquals(1d, anns[0].getScore(), 0.001d);
    assertEquals("my Score Description", anns[0].description);
    assertEquals("my Score", anns[0].label);
    anns = al.getSequenceAt(1).getAnnotation();
    assertEquals(1, anns.length);
    assertEquals(2d, anns[0].getScore(), 0.001d);
    assertEquals("my Score Description", anns[0].description);
    assertEquals("my Score", anns[0].label);
    anns = al.getSequenceAt(2).getAnnotation();
    assertEquals(1, anns.length);
    assertEquals(3d, anns[0].getScore(), 0.001d);
    anns = al.getSequenceAt(3).getAnnotation();
    assertEquals(1, anns.length);
    assertEquals(4d, anns[0].getScore(), 0.001d);
  }

  /**
   * Test with a description pattern that matches any string (or none), followed
   * by a 'number character' (0-9+.), followed by at least one separator
   * character, followed by at least one 'number character', then any trailing
   * characters.
   */
  @Test
  public void testGetScoresFromDescription_twoScores()
  {
    String regex = ".*([-0-9.+]+).+([-0-9.+]+).*";
    final int count = pp.getScoresFromDescription("my Score",
            "my Score Description", regex, true);
    System.out.println("Matched " + count + " for " + regex);
    assertEquals(3, count);

    /*
     * Seq1 has two score values parsed out
     */
    AlignmentAnnotation[] anns = al.getSequenceAt(0).getAnnotation();
    assertEquals(2, anns.length);
    assertEquals(1d, anns[0].getScore(), 0.001d);
    assertEquals("my Score Description", anns[0].description);
    assertEquals("my Score", anns[0].label);
    assertEquals(1d, anns[1].getScore(), 0.001d);
    assertEquals("my Score Description (column 1)", anns[1].description);
    assertEquals("my Score_1", anns[1].label);

    /*
     * Seq2 has no score parsed out (is this right?)
     */
    assertNull(al.getSequenceAt(1).getAnnotation());

    /*
     * Seq3 has two score values parsed out
     */
    // TODO parsed values (1.0 and 3.0) look wrong v description
    // would expect 2.0 and 0.1
    // undesired 'greedy' behaviour of regex?
    anns = al.getSequenceAt(2).getAnnotation();
    assertEquals(2, anns.length);
    assertEquals(1d, anns[0].getScore(), 0.001d);
    assertEquals("my Score Description", anns[0].description);
    assertEquals("my Score", anns[0].label);
    assertEquals(3d, anns[1].getScore(), 0.001d);
    assertEquals("my Score Description (column 1)", anns[1].description);
    assertEquals("my Score_1", anns[1].label);

    /*
     * Seq3 has two score values parsed out
     */
    // TODO parsed values (1.0 and 4.0) look wrong v description
    // would expect 3 and 0.01
    anns = al.getSequenceAt(3).getAnnotation();
    assertEquals(2, anns.length);
    assertEquals(1d, anns[0].getScore(), 0.001d);
    assertEquals("my Score Description", anns[0].description);
    assertEquals("my Score", anns[0].label);
    assertEquals(4d, anns[1].getScore(), 0.001d);
    assertEquals("my Score Description (column 1)", anns[1].description);
    assertEquals("my Score_1", anns[1].label);
  }
}